Puerto Rican Parrots and Potential Limitations of the Metapopulation Approach to Species Conservation MARCIA H. WILSON CAMERON B. KEPLER* U.S. Fish and WUdlife Service Patuxent Wildlife Research Center Laurel, MD 20708, U.S.A. NOEL F. R. SNYDER Parrot Programs, Wildlife Preservation Trust International P.O. Box 426 Portal, AZ 85632, U.S.A. SCOTT R. DERRICKSON Ornithology, Conservation Research Center National Zoological Park Front Royal, VA 22630, U.S.A. F. JOSH DEIN U.S. and WUdlife Service National Wildlife Health Research Center 6006 Schroeder Road Madison, Wl 53711, U.S.A. JAMES W. WILEY U.S. Fish and WUdlife Service Grambling Cooperative WUdlife Project Grambling State University P.O. Box 815 Grambling, LA 71245, U.S.A. JOSEPH M. WUNDERLE, JR. ARIEL E. LUGO Institute of Tropical Forestry Southern Forest Experiment Station USDA Forest Service Rio Piedras, PR 00928 DAVID L. GRAHAM Schubot Exotic Bird Health Center Texas Veterinary Medical Center Texas A & M University CoUege Station, TX 77843, U.S.A. WILLIAM D. TOONE San Diego W?d Animal Park 15500 San Pasqual Valley Road Escondido, CA 92027, U.SA. Abstract: Population viability analyses for a number of en- dangered species have incorporated a metapopulation ap- proach. The risk assessments of these viability analyses have indicated that some extant populations should be subdi- vided into numerous subgroups with exchange of individu- als among them in order to reduce the chance of catastrophic loss of the species. However, routine application of a policy of extensive subdivision may have detrimental consequences for certain endangered species. We examine the Puerto Rican Parrot asa case history in which this policy is ill-advised In 1989, a population viability analysis was conducted for the 'Mailing address: Southwest Research Group, School of Forest Re- sources, University of Georgia, Athens, GA 30602, USA Paper submitted March 16, 1992; revised manuscript accepted Jan- uary 19, 1993- 114 Las cotorras portornique?as y las limitaciones potenciales de la estrategia de metapoblaciones aplicada a la conservaci?n de especies Resumen: Los an?lisis de viabilidad de poblaciones hechos para un gran n?mero de especies en peligro de extinci?n ahora tienen un enfoque de metapoblaci?n Para reducir las probabilidades de p?rdidas catastr?ficas de individuos de una especie, poblaciones existentes se subdividen en nu- merosos subgrupos; ?sto se hace de acuerdo a los resultados de los an?lisis de riesgo que son parte integral del enfoque de an?lisis de viabilidad de poblaciones. Sin embargo, la sub- divisi?n cont?inua y rutinaria de poblaciones de especies en peligro de extinci?n puede ser detrimental para la sobre- vivencia de la especie. Examinamos al caso de la Cotorra Puertorrique?a d?nde este tipo de enfoque no es recomen- Conservation Biology, Pages 114-123 Volume 8, No. 1, March 1994 Wilson et ?. Metapopuhtiotts, Disease, and Conservation 115 parrot The document recommended subdivision of the ex- isting small captive flock into three groups. One of these captive flocks would consist of individuals transferred to a multi-species facility in the continental United States. Sub- sequently, individuals from this facility would be exchanged with the insular captive population(sJ and the relict wild flock For two reasons, implementation of this recommen- dation might have led to serious repercussions. First, this parrot, like many endangered species, has gone through a genetic bottleneck and may have a heightened susceptibility to disease. Multi-species facilities are a high-risk environ- ment favoring the transmission of pathogens, especially when the facilities are located outside the natural ranges of a particular species. Second the parrot is a K-selected species for which mate selection is idiosyncratic. This type of species often proves difficult to breed in captivity in small groups. Part of the problem in mate selection may be reduced by a policy allowing frequent transfers of individuals among fa- cilities, but such movements increase the chances of spread- ing disease in the metapopulation Thus, population viabil- ity analyses need to acknowledge that proliferation of captive subgroups accompanied by exchanges of individuals can in themselves carry substantial risks that must be weighed against the presumed benefits of subdivision dable. En el 1989, un an?lisis de viabilidad poblacional se efectu? para la Cotorra Puertorrique?a Como resultado del an?lisis se recomend? la subdivisi?n en tres grupos de la peque?a poblaci?n cautiva existente. Se recomend?, adem?s, transferir uno de los tres grupos a una facilidad con m?lti- ples especies en los Estados Unidos de Am?rica. Subsiguien- temente, individuos de esta facilidad ser?an intercambiados con las poblaciones cautivas en Puerto Rico y con la pobla- ci?n de cotorras a?n sobreviviendo bajo condiciones natu- rales. Hay dos razones por las cuales tales recomendaciones podr?an haber repercutido seriamente sobre la cotorra En primer lugar, esta especie, como muchas otras en peligro de extinci?n, ha atravesado por un cuello de botella gen?tico y como resultado de esto su suceptibilidad a enfermedades ha incrementado. Facilidades con m?ltiples especies son ambi- entes de alto riesgo para especies como la cotorra pues en ellos se favorece la transmisi?n de pat?genos, especialmente cuando los facilidades est?n localizadas lejos del ambiente natural de especies particulares. En segundo lugar, la co- torra es una especie tipo Yapara la cu?l la selecci?n de pare- jas es idiosincr?tica Este tipo de especie es generalmente dif?cil de propagar en cautiverio particularmente cuando las poblaciones son peque?as. Parte del problema de selecci?n de parejas se puede aliviar intercambiando frecuentemente individuos entre facilidades, pero tal estrategia aumenta el riesgo de transmisi?n de enfermedades entre las metapobla- c?ones. Por lo tanto, el an?lisis de viabilidad de poblaciones debe considerar que la proliferaci?n de subgrupos cautivos acompa?ados por el intercambio continuo de individuos en esos subgrupos, por su cuenta, conlleva riesgos sustanciales que se deben contraponer ante los presuntos beneficios de la subdivisi?n de las poblaciones. Introduction A primary objective of applied conservation biology is to minimize extinction rates of threatened species (Goodman 1987; Quinn & Hastings 1987). This orien- tation led to defining minimum viable populations and to incorporating the metapopulation concept as a frame of reference into species conservation (Shaflfer 1985; Simberlofif 1988). Gilpin and Soul? (1986) introduced the process known as population viability (sometimes vulnerability) analysis (PVA) to estimate minimum viable populations and to plot conservation strategies for endangered spe- cies. The PVAs produced to date have involved both intrinsic population characteristics (morphology, phys- iology, reproduction, disease resistance, mobility, be- havior, etc.) and environmental constraints (habitat quantity and quality, interacting species, disease vec- tors, etc.). These features may allow estimation of ex- tinction probabilities and development of strategies to minimize these probabilities. As noted by Shaffer (1990), PVA is an emerging technique of risk analysis that does not as yet have a well-developed methodology or widely accepted standards. The PVA process has been applied to about 46 species, including a variety of threatened vertebrates, invertebrates, and plants (U. S. Seal, personal communication). Recently, a metapopulation approach has been incor- porated into many PVAs (for example, the 1989 Florida Panther Viability Analysis and Species Survival Plan and the 1989 Puerto Rican Parrot PVA). For wild popula- tions, a metapopulation is a complex of interdependent subpopulations affected by recurrent extinctions and linked by recolonization flrom one or more large reser- voir populations (Murphy et al. 1990). For many endan- gered species, the PVA process has been expanded to include wild and captive subpopulations (Lacy et al. 1989; Seal & Lacy 1989). To reduce chances of catastrophic loss of all members of an endangered species and to counter genetic drift. Lacy (1987) recommended that managers subdivide a managed population into numerous breeding subgroups that exchange a few individuals per generation. For spe- cies with small total populations, the subdivisions can Conservation Biology Volume 8, No. 1, March 1994 116 Met?popula?ons, Disease, and Conservation W?son et al. involve very small numbers of individuals in each sub- population?^for example, as few as 12 Puerto Rican Par- rots (Lacy et al. 1989). The costs of dividing wild populations into small sub- populations involve an increased probability of extinc- tion due to demographic stochasticity and a possible loss of genetic variation. If members of each subpopu- lation reproduce, these costs can theoretically be miti- gated in captivity by regularly exchanging individuals among facilities or sites (in the wild). However, blanket application of a metapopulation approach to species conservation can be detrimental for certain endangered species, especially for species with a high susceptibility to diseases or with difiBlculty of breeding in small captive groups. In this report, we address these two factors for optimal management of endangered species in captivity, with special reference to the endangered Puerto Rican Parrot {Amazona vittata). Disease Disease can play a major role in the conservation of endangered species, as evidenced by a canine distemper epizootic which led to the extirpation of the black- footed ferret (Mustela nigripes) in the wild (Thorne & Williams 1988). Endangered species may often have heightened susceptibility to disease because of reduced genetic variation in small populations (O'Brien et al. 1985; Watkins et al. 1990) and because many endan- gered species (especially insular species) occur in iso- lated populations without previous exposure to certain pathogens (van Riper et al. 1986; Cooper 1989; Jenkins et al. 1989). This heightened susceptibility to disease is often apparent when endangered species are compared with closely related but nonendangered species. For ex- ample, the endangered Whooping Cranes (Grus amer- icana) at the Patuxent Wildlife Research Center have been much more susceptible than the Sandhill Cranes (G. canadensis) to coccidia and viral encephalitis (Dein et al. 1986). Snyder et al. (1985) reported nestling mortalities from herpesvirus infections of the endangered Mauritius Pink Pigeon {Columba mayeri) fostered under domes- tic pigeons (C livia) at the Rio Grande Zoo in Albu- querque. The foster pigeon flock apparently had been free of disease for several years. Four fertile Mauritius Pink Pigeon eggs were fostered to the surrogate birds. However, latently infected foster pigeons shed herpes- virus during the reproductive period, and each of the Mauritius Pink Pigeon hatchlings died suddenly at 5 to 11 days of age. No deaths among nestling domestic pi- geons in the foster flock were due to viral infection. The Mauritius Pink Pigeons seemed to have had no previous experience with this virus. The transfer of live animals or eggs to non-native areas always increases the risk of exposing immigrants to pathogens to which they are not adapted and the risk of spreading pathogens from one area to another (Ashton & Cooper 1989; Gaskin 1989). For example, O'Brien et al. (1985) described the spread of a coronavirus throughout a cheetah {Acinonyx jubatus) colony. These authors attributed the cause to a clinically healthy female transferred to Wildlife Safari. Despite intensive therapy, this female and 17 other cheetahs died. O'Brien et al. (1985) suggested that a dearth of genetic variation rendered this colony extremely susceptible. Gaskin (1989) suggested that congregation of a di- verse array of species from all around the world in quar- antine facilities and large, multi-species holdings in zoos, private collections, and the pet trade is one reason cap- tive psittacines are plagued with so many diseases. These multi-species congregations have provided en- hanced opportunities for the transmission of infectious agents. Quarantine, a common practice in the transfer of live animals to new locations, does not guarantee that an individual is free of diseases. Many disease agents are not necessarily fatal on contact but can remain latent in carriers and be shed intermittently, particularly under stress. Carriers can harbor certain diseases for their en- tire lives and may transmit infections to their oflfepring. The most disturbing fact is that carriers of some disease agents can be asymptomatic. Such is the case with a psittacine herpesvirus known as Pacheco's disease (Gaskin 1989), and with a number of other avian dis- eases. Shima and Osborn (1989) identified a pair of newly acquired Red-Collared Lorikeets (Trichoglossus haema- todus rubritorquis) as the "point source" of an eporni- tic of Salmonella typhimurium in the San Diego Zoo collection of lories and lorikeets. Seven birds, represent- ing six species, died. The originally infected pair was imported from an Australian zoo, went through standard federal quarantine, and then a 30-day quarantine at a holding facility at the San Diego Wild Animal Park be- fore being introduced into the collection at the zoo. Routine screening failed to identify the problem during quarantine. The purpose of most zoos is to display animals, and many species are housed in close proximity for viewing by the public. To replace animals that die and to diver- sify collections for public viewing, zoos continually ac- quire new individuals and species from around the world. In part, because these animals often come from aggregations of species in quarantine or from other fa- cilities, zoos may be high-risk environments that favor transmission of pathogens. Not surprisingly, many epi- zootics have been documented in zoological facilities (see Montali et al. 1975?; May 1988; Thorne & Williams 1988; Cooper 1989; Derrickson & Snyder 1992). For threatened birds that might be highly susceptible to diseases and yet are in need of captive propagation. Conservation Biology Volume 8, No. 1, March 1994 Wilson et al. Metapopulations, Disease, and Conservation 111 Derrickson and Snyder (1992) recommended the fol- lowing principles: ( 1 ) Facilities should, whenever possible, be located in the species' natural range. (2) The species should be housed in two or three geo- graphically isolated, single-species facilities. (3) The facilities should be staffed by individuals who are not simultaneously caring for other captive birds. (4) As much as possible, facilities should be in areas free from arthropod vectors and feral populations of exotic bfrds. (5) Established husbandry protocols should empha- size disease prevention. These precautions do not preclude the involvement of zoological institutions in breeding of endangered spe- cies that are susceptible to diseases, but they emphasize proper locations and careftiUy controlled conditions. Likewise, Kennedy-Stoskopf (1989) pointed out that an appropriate environment and good husbandry practices are essential to limiting infectious diseases in captive situations. Recovery programs for endangered species increas- ingly include augmenting or reestablishing wild popu- lations with captive individuals or with wild, translo- cated conspecifics. If either type of individual harbors pathogenic organisms, relict wild populations can be severely jeopardized. For example, many reintroduced populations of Wild Turkeys {Meleagris gallopavo} in the U.S. are infected with Plasmodtum kempt, a hema- tozoan parasite that appears to have originated from the translocation of infected birds (Castle & Christensen 1990). In a like manner, there is convincing evidence that a movement of rabies up the East Coast was initi- ated by the translocations of infected raccoons from endemic areas in Florida and Georgia (Nettles et al. 1979). Similarly, the upper respiratory disease in wild desert tortoises (JKerobates agassizii} is believed to stem from a probable release of infected pet (or captive- held) tortoises (Anonymous 1989; Jacobson & Gaskin 1990). This disease now threatens the species. Infectious and parasitic diseases can cause premature death in wild populations (Scott 1988; Cooper 1989). Fatal diseases are notoriously difficult to detect in wild animal populations because sick and dead individuals commonly go unrecovered, in part because they are taken quickly by predators and scavengers (Balcomb 1986; Scott 1988; Unz et al. 1991). Even when a caus- ative agent can be identified in a diseased wild popula- tion, treatment is often impractical, particularly if spe- cies are difficult to catch or easily stressed by handling or if practical means of treatment are unknown. Furthermore, disease is most likely to affect a popu- lation without previous contact with the pathogen. For example, as noted by Cooper (1989), if the herpesvLrus that afflicted the Pink Pigeons in the Rio Grande Zoo had been introduced into the wild Pink Pigeon popula- tion in Mauritius, results could have been disastrous. With good reason, Ritchie et al. (1989) stressed that, because of the widespread movement of birds in the pet trade, the risks of introducing highly virulent viruses into wild populations of endangered psittacine species are substantial. Besides causing premature mortality, introduction of infected captive animals into the wild can also reduce reproduction of wild populations. For example, Hudson (1986) found that reproduction in Red Grouse {Lago- pus scottus) can be reduced by the nematode {Tricho- strongylus tenuis). Similarly, Saumier et al. (1986) found that captive American Kestrels {Falco sparverius} infected with the parasite Trichinella pseudospiralis had severely reduced reproductive success compared with control birds. Infected pairs initiated laying later, laid smaller clutches, and hatched fewer eggs than con- trols. Identifying an infection, rather than some other factor, as the actual cause of low reproductive success in a wild population is difficult. Because some avian pathogens can be transmitted in eggs, one cannot avoid transport of pathogens by limit- ing movements of animals to the egg stage. The bacte- rium Salmonella pullorum can spread from a parent flock via the egg to chicks by true transovarian vertical transmission (Ashton & Cooper 1989). Other avian pathogens can also enter eggs as they cool following surface contamination of the shell (Ashton & Cooper 1989). Vertical transmission is also known to occur with avian leukosis virus, reticuloendotheiiosis, adenovi- ruses, and Mycoplasma spp. (Reece 1989). ff an in- fected embryo does not die, infection of newly hatched chicks may occur and spread throughout a captive flock. Because novel pathogens or their vectors can be spread with increasing ease in the modern world. Coo- per (1989) made an urgent plea for minimizing the spread of infectious diseases from one locality to an- other. As Ashton and Cooper (1989) pointed out, ex- clusion of an avian pathogen is much easier than its elimination. Elimination of a disease requires that it is well de- fined; that it can be easily diagnosed or detectable by a simple, cheap, and rapidly applied test (which is often not available); and that it can be eradicated. But under- standing of many diseases, especially avian diseases, is still far from perfect. Part of the difficulty stems from the sheer number of species from all over the world that are involved. Host ranges are unknown for some diseases such as psittacine beak and feather disease, and caus- ative agents are stiU unknown for others such as psitta- cine proventricular dilatation syndrome (parrot wasting disease) and internal papillomatous disease. Of the many infectious agents, Gaskin (1989) reported that vi- ruses are the most troublesome mainly because they can Conservation Biology Volume 8, No. 1. March 1994 118 Metapopulations, Disease, and Conservation Wilson et al. occur in epornitic proportions with little possibility of chemotherapeutic intervention. Once certain pathogens become established in a fa- cility, they can be extremely difficult to control or elim- inate. One such disease that has repeatedly plagued zoo- logical and private collections and the poultry industry is avian tuberculosis. This disease affects a wide variety of avian taxa, and it is readily spread by fecal contami- nation of substrates, feed, or water sources, and possibly also vertically through egg contamination (Montali et al. 1978). Genetic factors can also play a role in transmis- sion (Cromie et al. 1991 ). Control or eradication of this disease is particularly difficult because the causative ba- cillus, Mycobacterium avium, can remain viable in the soil for years, and because reliable methods to diagnose and treat the disease in live animals remain unavailable. Traditional eradication includes the elimination of tlie entire collection and either the replacement of existing substrates, the destruction of the entire facility, or movement of the facility to a new site (Montali et al. 1975fc; Beehler 1990). Such measures obviously entail severe financial, biological, and ethical costs (particu- larly when dealing with endangered and threatened spe- cies), and do not preclude accidental reinfection. Con- trolling the spread of avian tuberculosis among facilities or between captive and wild populations clearly will not be possible until efficient and accurate diagnostic tech- niques become available to detect the disease in live animals. Unfortunately, avian tuberculosis has been de- tected post-mortem in several endangered species, and this insidious disease represents a substantial threat to many ongoing breeding and reintroduction programs. Polyomaviruses, which are significant pathogens of psittacines, are also resistant to inactivation. For exam- ple, Gough (1989) reported on an aviary in which an outbreak of a Budgerigar fledgling disease occurred. Af- ter the initial outbreak, breeding operations were cur- tailed for eight months, but young continued to die when breeding was resumed. The entire aviary was de- populated, fumigated with formaldehyde gas, and left vacant for seven months. Then new breeding stock was purchased from an aviary presumably free of the disease. Despite these massive efforts to eliminate the disease, it soon broke out again. Likewise, Thorne and Williams (1988) reported that canine parvoviral enteritis, a very contagious infectious virus of canids, is resistant to environmental inactiva- tion. Apparently carrier wolves may have been the source of this virus, which caused mortalities among maned wolves {Chysocyon brachyurus) both at the San Antonio Zoo (Fletcher et al. 1979) and at the National Zoological Park (Mann et al. 1980). In summary, because of the slow-acting but neverthe- less lethal or debilitating nature of many pathogens, be- cause of the difficulties in detecting the presence of many pathogens in asymptomatic carriers, because of the widespread movement of animals carrying patho- gens through captive-breeding collections around the world, because of the heightened susceptibilities of many endangered species to disease, and because of se- vere risks of contamination of wild populations of en- dangered species by infected individuals from captivity, it is questionable that extensive subdivision of captive populations of endangered species, especially in multi- species facilities, represents a risk-reduction strategy rather than a risk-enhancement strategy. This is espe- cially true when movements between subpopulations are routinely required to meet genetic and demographic management objectives. Presumably, many of the risks of disease can be greatly reduced if subdivision of cap- tive populations is limited to single-species facilities within the natural range of the endangered species, and if personnel serving the facilities practice rigorous dis- ease prevention measures, including avoidance of con- tact with other animals that may be harboring disease agents. Poor Reproduction of Small Populations Housed in Captivity The productivity of many species in captivity is affected strongly by population size. In many colonial species, social facilitation is a necessary element for reproduc- tion. For this reason, flamingos and many other colonial waterbirds such as storks, ibises, and spoonbills, usually do not breed in captivity in small groups (Kear & Palmes 1980; Luthin et al. 1985). For many endangered species, obtaining adequate lev- els of reproduction in captivity poses difficulties. For species with idiosyncratic mate selection, subdivision of an already small population reduces the number of pos- sible pairings at any one site and therefore the overall productivity of the species. In many large psittacines, reproduction occurs in only a small fraction of a captive population, and success in captive breeding seems to depend on providing birds with an ample array of potential mates. Even with nu- merous trials with potential mates, many birds remain nonreproductive, presumably because of their failure to find compatible partners among the available choices. The psittacine literature is rife with examples of the contortions aviculturists have been through to obtain compatible pairs. Even in psittacines, such as Cockatiels {Nymphtcus hollandicus), that breed relatively freely in captivity, forced pairing commonly results in much less successful reproduction than free mate choice (Ya- mamoto et al. 1989). The spontaneous choice of mates seems to be crucial for successful reproduction in many avian species, such as Canvasbacks {Aythya valisineria) (Bluhm 1985) and Whooping Cranes (Kepler 1978). With species in which formation of compatible pairs is especially difficult, extensive subdivision of captive Conservation Biology Volume 8, No. 1, March 1994 Wilson et ?. Metapopuktiotts, Diseuse, and Conservation 119 populations can pose significant impediments to secur- ing sustained reproduction and meeting genetic man- agement objectives. Both of these problems can be mit- igated only by extensive and sometimes expensive transfers of individuals among facilities. Such move- ments, however, can greatly increase the chance of spreading disease among subpopulations and can in- crease the vulnerability of the entire metapopulation to disease. Thus, the perceived reduction of risk from splitting a captive population to prevent loss of an entire genome to some site-specific catastrophe, such as fire, disease, tornados, or local wars, must be balanced against the potential increase in disease exposure and decrease in reproductive potential that may result from subdivision. Viable or optimal group size varies greatly from species to species. A Case Study: The Puerto Rican Parrot The Puerto Rican Parrot was abundant and widespread on the islands of Puerto Rico, Culebra, and Vieques. By the late 1960s, it was reduced to a tiny remnant popu- lation in the Luquillo Experimental Forest, also known as the Caribbean National Forest. This reduction was caused by a variety of factors, including habitat destruc- tion, capture for pets, and killing for food and protection of crops (Snyder et al. 1987). The wild population reached a low point of 13 birds at the beginning of the 1975 breeding season. Since then it has recovered at a low to moderate rate, reaching a high of 45-47 individ- uals after the 1989 breeding season. This increase was a result of identification and reduction of several limiting factors afifecting the wild population, and of fostering captive-reared birds into wild nests. Meanwhile, a cap- tive population was started in the early 1970s largely through removal of eggs from wild nests, and reached a total of 58 individuals by January 1993 (A. Valido, per- sonal communication). Although successful captive breeding has been achieved in all years since 1979, the number of breeding pairs has remained low in both wild and captive flocks, averaging about four pairs in each flock. The captive flock is in a single facility within the natural range of the wild population in the Luquillo Experimental Forest. With the exception of captive Hispaniolan Parrots (Am- azona ventralis) brought in for surrogate studies, the aviary has been essentially closed to importation of other avian species. No significant disease problems have occurred at the facility in the 20 years of its exis- tence. A second aviary for the species, about 100 km distant, has recently been completed. The site of this second aviary is in the Rio Abajo Commonwealth Forest, which has been acknowledged to be the best location to rees- tablish a second wild population of the species (U.S. Fish and Wildlife Service 1987). In June of 1989, the U.S. Fish and Wildlife Service hosted the Captive Breeding Specialist Group of the In- ternational Union for the Conservation of Nature and Natural Resources in a PVA on the Puerto Rican Parrot. The PVA report made several recommendations for management of wild and captive flocks, but most nota- bly it advocated the near-term splitting of the captive flock into three parts: the existing Luquillo Aviary in the Luquillo Experimental Forest, the second aviary in Rio Abajo, and a third aviary at some mainland zoo in the United States (Lacy et al. 1989). Ultimately, the PVA recommended the establishment of some 10 captive flocks in various locations, including zoos. These rec- ommendations were based on the application of the metapopulation concept. The principal justifications offered for sending birds to a mainland zoo were: ( 1 ) enhanced reduction of risks from catastrophe, and (2) enhancement of captive re- production utilizing the resources and expertise at a mainland zoo. The 12 birds recommended for shipment to a mainland zoo were all birds that had not bred as yet in captivity. The primary threat to both wild and captive popula- tions of the Puerto Rican Parrot has long been perceived to be a major hurricane disaster. By 1989, the last wild habitat of the species had been spared any major hurri- cane for over 50 years. The effects of a major storm were difficult to predict but were perceived to be sig- nificant for both wild and captive flocks, and a consen- sus has long existed that establishment of a second cap- tive flock and a second wild flock as soon as practical would be best. Consensus on the advisability of starting a third cap- tive flock, especially one in a mainland zoo location, has been more elusive. Debate on this issue occupied the greater part of 1990 and centered on the two major issues raised in the earlier part of this report: disease risks and projected effects on reproduction. The Puerto Rican Parrot is perhaps a classic example of a species for which high disease susceptibility is predicted. As a member of the Psittacidae, it is a potential host of more than 30 known diseases and potential disease syn- dromes. Moreover, it is a highly endangered, insular spe- cies that has been in a genetic bottleneck for nearly 25 years, a period during which the number of breeding pairs has been low, very likely resulting in a high degree of relatedness and reduced genetic diversity among sur- viving individuals. Opponents of the move of birds to a mainland zoo pointed out that a variety of psittacine diseases are very difficult to detect, yet are widespread in mainland zoos and could potentially be spread to captive and wild flocks in Puerto Rico if interchanges of birds were car- Conservation Biology Volume 8, No. 1, March 1994 120 Metapopulations, Disease, and Conservation Wilson et al. ried out. Conversely, if interchanges were not carried out, it was dififlcult to see how a mainland zoo flock would have any conservation value for the existing wild population or for a prospective new wild population. Of special concern, the mainland zoo recommended to host Puerto Rican Parrots has had recurrent cases of avian tuberculosis, a disease that is very difficult to de- tect in carriers and extremely difficult to eradicate from a facility. Furthermore, this facility maintained a diverse, multi-species psittacine collection and, as a result, the risk of Puerto Rican Parrots contacting other diseases in this facility appeared to be substantial. Opponents of the move suggested that it would be wiser to export the expertise of a mainland zoo to facilities in Puerto Rico rather than to move birds to such a mainland environ- ment. This position was ultimately upheld by the U.S. Fish and Wildlife Service as a result of consultations with its own veterinary experts. Of equal concern were the potential effects on repro- duction from subdividing the captive flock three ways. The captive flock had never supported more than six successful breeding pairs, despite major efibrts to estab- lish compatible pairings and despite the fact that surro- gate Hispaniolan Parrots had been bred much more suc- cessfully at the same facility, using the same techniques. Production of captive young Puerto Rican Parrots has rarely exceeded six individuals in a year, despite exper- imentation with a variety of husbandry techniques. What would be the projected effects of splitting the existing captive flock three ways? Clearly, the choices for trying new pairings or allowing maturing birds to select mates would be greatly reduced for each facility from what is now possible. This could conceivably re- sult in a failure to achieve any successful new pairings in any facility. This expectation is based on a track record of needing more than a dozen individuals to form each new productive pair. If, indeed, the major problem in captive reproduction of this species is achieving com- patible pairs, a three-way split of the existing captive population could have severe consequences. Other possible causes of low captive production of the Puerto Rican Parrot likely exist, but the potential importance of these other causes must always be viewed against the consistently successful captive pro- duction of Hispaniolan Parrots at the same facility using the same diet and techniques as with the Puerto Rican Parrot. In any event, a case has yet to be made that solutions to the low captive reproduction of Puerto Ri- can Parrots can be found any more easily in a mainland environment than within the natural range of the spe- cies. Almost as if to test the strength of conservation strat- egies under debate, Puerto Rico was directly hit by a major storm. Hurricane Hugo, in September of 1989. This hurricane caused massive damage to the remnant habitat of the Puerto Rican Parrot. The parrot aviary, however, survived almost completely unscathed, and with no loss of birds. The wild population was reduced from 45?47 birds to about 25 birds. Interestingly, parrot breeding resumed in 1990. In 1991, a record six pairs of parrots bred in the wild, the highest total since the 1950s. The next year, these six pairs fledged a record of 11 young, and the post-breeding population estimate was between 34 and 37 parrots (A. Valido, personal communication). It presently appears that chances for a resumption of sustained recovery of the wild population are relatively good. Evidently, the amount of risk attrib- uted to a major storm can be overestimated, to both captive and wild populations. How many subpopulations of the Puerto Rican Par- rots are needed at the present time? Clearly the answer to this question is crucially dependent on how many birds are available and on how many birds constitute a critical mass for successful reproduction. Splitting either the existing wild flock or the existing captive flock car- ries penalties to these flocks that must be weighed against the benefits. The existing two subpopulations have just come through a major environmental disaster in relatively good shape, which should perhaps temper enthusiasm for splitting them up. Had the wild popula- tion been only half as large at the time the storm hit, things could look much worse now than they do. In our view, the wild flock has shown a surprisingly encouraging resilience and ability to increase over the past two decades (at a rate very similar to the recovery of the wild population of Whooping Cranes). Now it has also shown an encouraging survival through a major storm. We believe that preserving the integrity of this population represents the highest priority in future ef- forts to preserve the species. In view of the long history of relatively poor perfor- mance of the species in captivity, conservation efforts should not be skewed toward large-scale proliferation of captive populations; instead, the fruits of captive pro- duction in the near term should be used largely to bol- ster the existing wild population. Once that population reaches perhaps 70 to 100 individuals, it may obtain a fairly comfortable status with respect to future hurri- canes, and it will become important to begin the estab- lishment of a second wild flock. Meanwhile, we believe that establishment of a second captive flock in Rio Abajo represents a reasonable strat- egy, but it should not take place at the expense of the productivity of the existing captive or wild subpopula- tions. This second aviary should be established with cur- rent nonbreeders and juveniles from the existing cap- tive flock. No attempts to establish a second wild flock should be made until the existing wild flock exceeds 70 birds, because such efforts would conflict with bolster- ing the extant wild flock as quickly as possible while it Conservation Biology Volume 8, No. 1, March 1994 Wilson et ai. Metapopuiations, Disease, and Conservation 121 is still at a critically low level. We see no clear advantage in establishing a third captive flock at the present time. Conclusions The overriding goal of securing wild populations can sometimes be forgotten in the recent climate of enthu- siasm for captive breeding. The primary rationale usu- ally offered for establishing multiple captive populations is one of risk reduction from catastrophe. Unfortunately, it is not often acknowledged that proliferation of captive populations itself carries significant risks, especially with respect to disease vulnerability and reproductive penalties. The recent case history of the critically endangered Puerto Rican Parrot indicates that even major catastro- phes can have relatively limited effects on some popu- lations. This experience underscores the need for incor- porating more rigorous information on the species' ecology into genetic models and into concepts of metapopuiations. Species and populations have resil- iency mechanisms that may be as important as genetic make-up in their day-to-day survival. Therefore, deci- sions on their management must balance ecological and genetic arguments. The recovery potential of some species may also de- pend on critical-mass effects that argue against oversplit- ting populations. When subdivision involves moving en- dangered species into non-native areas, it places them in close proximity to potential vectors of new diseases. The risks involved cannot be rigorously quantified in magnitude, but they appear to be well worth avoiding when alternatives exist. The consequences of mistakes can easily prove fatal to species. We wish to make perfectly clear that we do not op- pose an important role for zoological institutions in the conservation of endangered species, and in fact we strongly support such a role. Much remains to be im- proved, however, regarding the precautions taken in breeding endangered species in captivity, particularly in disease prevention and control. Attempts to house spe- cies that are arguably sensitive to disease in ex situ multi-species environments are generally unwise, and the should be replaced by a preference for breeding endangered species in situ in closed single-species facil- ities operating under rigorous disease-prevention pro- tocols. While this approach is in some cases more ex- pensive, such is not always the case, especially considering the costs of routine international transport in ex situ approaches. Furthermore, the benefits of in situ programs that involve further communities in the conservation of their native fauna can be as substantial as the benefits in reducing unnecessary risks of expo- sure to disease. Acknowledgments We thank M. Friend, W. Barrow, E. Rockwell, and two anonymous reviewers for comments on the manuscript. Literature Cited Anonymous. 1989. Emergency action taken to protect the des- ert tortoise. Endangered Species Teclinical Bulletin 14:1, 5-6. Asliton, W. L. G., and J. E. Cooper. 1989. Exclusion, elimination and control of avian pathogens. Pages 31?38 in J. E. Cooper, editor. Disease and threatened birds. Technical publication No. 10. Cambridge, England. Balcomb, R. 1986. Songbird carcasses disappear rapidly from agricultural fields. Auk 103:817-820. Beehler, B. A. 1990. Management o?Mycobacterium avium in a mixed species aviary. Pages 125?129 in Proceedings of the American Association of Zoo Veterinarians. Bluhm, C. K. 1985. Mate preferences and mating patterns of Canvasbacks {Aythya valisineria). Ornithological Mono- graphs 37:45-56. Castle, M. D., and B. M. Christensen. 1990. Hematozoa of Wild Turkeys from the midwestern United States: Translocation of wild turkeys and its potential role in the introduction oiPlas- modium kempi Journal of Wildlife Diseases 26:180-185. Cooper, J. E. 1989. The role of pathogens in tlireatened pop- ulations: An historical review. Pages 51-61 in J. E. Cooper, editor. Disease and threatened birds. Technical publication No. 10. , Cambridge, England. Cromie, R. L, M.J. Brown, D.J. Price, and J. L. Stanford. 1991. Susceptibility of captive waterfowl to avian tuberculosis. Tu- bercle 72:105-109. Dein, F.J., J.W. Carpenter, G.G. Clark, R.J. Montali, C. L Crabbs, T. F. Tsai, and D. E. Docherty. 1986. Mortality of cap- tive Whooping Cranes caused by eastern equine encephalitis virus. Journal of the American Veterinary Medical Association 189:1006-1010. Derrickson, S. R., and N. F. R. Snyder. 1992. Potentials and lim- its of captive breeding in parrot conservation. Pages 133?163 in S. R Beissinger and N. F. R. Snyder, editors. New World par- rots in crisis: Solutions from conservation biology. Smithso- nian Institution Press, Washington, D.C. Fletcher, K. C, A. K. Eugster, R. E. Schmidt, and G. B. Hubbard. 1979. Parvovirus infection in maned wolves. Journal of the American Veterinary Medical Association 175:897?900. Gasldn, J. M. 1989. Psittacine viral diseases: A perspective. Journal of Zoo and WUdlife Medicine 20(3):249-264. Gilpin, M. E., and M. E. Soul?. 1986. Minimum viable popula- tions: Processes of species extinction. Pages 19?34 in M. E. Soul?, editor. Conservation biology: The science of scarcity and diversity. Sinauer Associates, Sunderland, Massachusetts. Conservation Biology Volume 8, No. 1, March 1994 122 Metapopuhtions, Disease, and Conservation W?son et ?. Goodman, D. 1987. How do any species persist? Lessons for conservation biology. Conservation Biology l(l):59-62. Gough, J. F. 1989. Outbreaks of budgerigar fledgling disease in three aviaries in Ontario. Canadian Veterinary Journal 30:672? 674. Hudson, P. J. 1986. The effect of a parasitic nematode on the breeding production of Red Grouse. Journal of Animal Ecology 55:85-92. Jacobson, E. R, and J. M. Gaskin. 1990. Clinicopathologic in- vestigations on an upper respiratory disease of free-ranging desert tortoises, Xerobates agassizii. U.S. Contract report. Bu- reau of Land Management, Department of the Interior, River- side, California. Jenkins, C. D., S. A. Temple, C. van Riper, and W. R. Hansen. 1989. Disease-related aspects of conserving the endangered Hawaiian Crow. Pages 77?87 in J. E. Cooper, editor. Disease and threatened birds. Technical publication No. 10. Cambridge, England. Kear, J., and P. Palmes. 1980. Andean and James Flamingos. International Zoo Yearbook 20:17-23. Kennedy-Stoskopf, S. 1989. Infectious diseases: Cause or ef- fects? Journal of Zoo and Wildlife Medicine 20(3):247-248. Kepler, C. B. 1978. Captive propagation of Whooping Cranes: A behavioral approach. Pages 231-241 in S. A. Temple, editor. Endangered birds: Management techniques for preserving threatened species. University of Wisconsin Press, Madison, Wisconsin. Lacy, R. C. 1987. Loss of genetic diversity from managed pop- ulations: Interacting effects of drift, mutation, immigration, se- lection, and population subdivision. Conservation Biology 1(2):143-158. Lacy, R. C, N. R. Flesness, and U. S. Seal. 1989. Puerto Rican Parrot (Amazona vittata) population viability analysis and recommendations. Captive Breeding Specialist Group, Apple Valley, Minnesota. Linz, G. M., J. E. Davis, Jr., R. M. Engeman, D. L. Otis, and M. L. Avery. 1991. Estimating survival of bird carcasses in cattail marshes. WildUfe Society Bulletin 19:195-199. Luthin, C. S., G. W. Archibald, L. Hartman, C. M. Mirande, and S. Swengel. 1985. Captive breeding of endangered cranes, storks, ibises, and spoonbills. International Zoo Yearbook 24/25:25? 39. Mann, P.C., M. Bush, M.J. G. Appel, B.A. Beehler, and R.J. Montali, 1980. Canine parvovirus infection in South American canids. Journal of the American Veterinary Medical Associa- tion 177:779-783. May, R. M. 1988. Conservation and disease. Conservation Bi- ology 2(l):28-30. Montali, R. J., M. Bush, C. O. Thoen, and E. Smith. 1976&. Tu- berculosis in captive exotic birds. Journal of the American Veterinary Medical Association l69(9):920-927. Montali, R. J. M. Bush, and E. E. Smith. 1978. Pathology of tu- berculosis in captive exotic birds. Pages 209-215 in Proceed- ings: Mycobacterial infections of zoo animals symposium. Smithsonian Institution Press, Washington, D.C. Murphy, D. D., K. E. Freas, and S. B. Weiss. 1990. An environ- ment-metapopulation approach to population viability analysis for a threatened invertebrate. Conservation Biology 4(l):4l- 51. Nettles, V. F., J. H. Shadduck, R. K. Sikes, and C. R. Reyes. 1979. Rabies in translocated raccoons. AJPH 69(6):601-602. O'Brien, S.J., M.E. Roelke, L. Marker, A. Newman, C.A. Win- kler, D. Meltzer, L. Colly, J. F. Evermann, M. Bush, and D. E. Wildt. 1985. Genetic basis for species vulnerability in the cheetah. Science 227:1428-1434. Quinn, J. F., and A. Hastings. 1987. Extinction in subdivided habitats. Conservation Biology l(3):198-208. Reece, R, L. 1989. Avian pathogens: Their biology and methods of spread. Pages 1?23 in J. E. Cooper, editor. Disease and threatened birds. Technical publication No. 10. , Cam- bridge, England. Ritchie, B. W., F. D. Niagro, P. D. Lukert, K. S. Latimer, W. L. Steffens, and N. Pritchard. 1989. A review of psittacine beak and feather disease. Journal of the Association of Avian Veter- inarians 3(3):143-148. Saumier, M. D., M. E. Rau, and D. M. Bird. 1986. The effect of Trichinella pseudospiralis infection on the reproductive suc- cess of captive American Kestrels (Falco sparveritds). Cana- dian Journal of Zoology 64:212 3-2125. Scott, M. E. 1988. The impact of infection and disease on ani- mal populations: Implications for conservation biology. Con- servation Biology 2(1):40?56. Seal, U. S., and R. C. Lacy. 1989. Florida panther (Felis concolor coryi) viability analysis and species survival plan. Captive Breeding Specialist Group, Apple Valley, Minnesota. Shaffer, M. L. 1985. The metapopulation and species conserva- tion: The special case of the Northern Spotted Owl. Pages 86-99 in R. J. Gutierrez and A. B. Carey, editors. Ecology and management of the Spotted Owl in the Pacific Northwest. General technical report PNW-185. U.S. Department of Agri- culture?Forest Service, Shaffer, M. L. 1990. Population viability analysis. Conservation Biology 4(l):39-40. Montali, R.J., M. Bush, and G.A. Greenwell. 1976? An epor- nitic of duck viral enteritis in a zoological park. Journal of the American Veterinary Medical Association l69(9):954-958. Shima, A. L, and K. G. Osborn. 1989. An epornitic of Salmo- nella typhimurium. in a collection of lories and lorikeets. Journal of Zoo and WUdlife Medicine 20(3):373-376. Conservation Biology Volume 8, No. 1, March 1994 Wilson et al. Metapopulations, Disease, and Conservation 123 Simberloflf, D. 1988. The contribution of population and com- munity biology to conservation science. Annual Review of Ecology and Systematics 19:473?511. Snyder, B., J. Thilsted, B. Burgess, and M. Richard. 1985. Pigeon herpesvirus mortalities in foster reared Mauritius Pink Pi- geons. Pages 69?70 in Proceedings of the American Associa- tion of Zoo Veterinarians. Snyder, N. F. R., J. W. WUey, and C. B. Kepler. 1987. The par- rots of LuquiUo: Natural history and conservation of the Puerto Rican Parrot. Western Foundation of Vertebrate Zoology, Los Angeles, California. Thorne, E. T., and E. S. Williams. 1988. Disease and endangered species: The black-footed ferret as a recent example. Conser- vation Biology 2(l):66-74. U.S. Fish and Wildlife Service. 1987. Recovery plan for the Puerto Rican Parrot, Amazona vittata. U.S. Fish and Wildlife Service, Atlanta, Georgia. van Riper, C, S. G. van Riper, M. L Goflf, and M. Laird. 1986. The epizootiology and ecological significance of malaria in Hawaiian land birds. Ecological Monographs 56:327-344. Watkins, D. L, N. L. Letvin, A. L. Hughes, and T. F. Fedder. 1990. Molecular cloning of cDNAs that encode MHC class I molecules from a New World primate {Saguinus oedipus): Natural selection acts at positions that may affect peptide pre- sentation to T cells. Journal of Immunology 144:1136-1143. Yamamoto, J. T., K. M. Shields, J. R. Millam, T. E. Roudybush, and C. R. Grau. 1989. Reproductive activity of force-paired Cockatiels {Nymphicus foollandicus). Auk 106:86-93. Conservation Biology Volume 8, No. 1, March 1994